[secdir] Secdir review of draft-ietf-rmt-bb-lct-revised-09

[Catherine Meadows <[e-mail]>]

I have reviewed this document as part of the security directorate's
ongoing effort to review all IETF documents being processed by the
IESG.  These comments were written primarily for the benefit of the
security area directors.  Document editors and WG chairs should treat
these comments just like any other last call comments.

This paper concerns the Layered Coding Transport building block.  LCT  
provides
support for massively scalable protocols using the IP multicast  
network service.
LCT supports sessions consisting or multiple channels, all with the  
same sender, but
many different receivers.  This makes it compatible with many  
massively scalable congestion
control protocols that use this structure.

When we are dealing with massively scalable congestion control  
protocols, the question
of denial of service naturally comes to mind.  Thus the security  
considerations section
rightly mainly concerns itself with that.  However, it left me with  
several questions.

1.  The introduction says that "Protocol Instantiations that use the  
LCT building block MUST
address the security requirements described in the following  
sections."   But there are no MUSTs
or MUST NOTs in the following sections.  Indeed,sections 8.1 and 8.2  
don't contain any recommendations
at all; they simply identify vulnerabilities (which could, however, be  
addressed by authentication).  I would
suggest either toning down the introduction ("MUST address the  
potential vulnerabilities" instead of "MUST address
the requirements") or beef up the specific sections.

2.  The sections themselves have a kind of piecemeal feel about them,  
addressing specific potential attacks, but
without giving a feeling that everything that is covered.  It might be  
a better idea to describe what services security mechanisms
could provide (e.g. authentication, confidentiality) describe the  
various each service offered by LCT and its vulnerabilities,
and then what security services would address.  I suspect, given the  
vulnerabilities that the authors have already described, it
would mostly boil down to identifying which LCT data should or must be  
authenticated.


Catherine Meadows
Naval Research Laboratory
Code 5543
4555 Overlook Ave., S.W.
Washington DC, 20375
phone: [phone]
fax: [phone]
email: [e-mail]